best-of

Compare two git worktrees against project guidelines and quality criteria.

2|Updated Mar 18, 2026
One-click install
npx skills add https://github.com/pmatos/skills --skill best-of
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: best-of
Source: https://github.com/pmatos/skills/tree/main/best-of
Command: npx skills add https://github.com/pmatos/skills --skill best-of

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill helps engineers and reviewers make informed judgments when faced with two code worktrees by performing a structured, evidence-based comparison against software engineering best practices and project contribution guidelines. It delivers a formal verdict on which solution is better and why.

Core Features & Use Cases

  • Side-by-side worktree comparison across correctness, design, testing, and adherence to project guidelines.
  • Structured evaluation workflow inspired by multi-agent analysis to surface strengths, weaknesses, and trade-offs.
  • Generated verdict and scorecard to guide decision making in code review, refactoring, or architecture discussions.

Quick Start

Provide two absolute paths to the worktrees you want to compare and the system will generate a detailed, evidence-based verdict.

Frequently Asked Questions about best-of

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I compare two git worktrees to determine which code solution is better?

To compare two git worktrees, provide their absolute paths and the system analyzes surface changes against project contribution guidelines, producing a structured scorecard and verdict across correctness, design, testing, and compliance.

What is structured code review evaluation and how does it work for multiple implementations?

Structured code review evaluation identifies and compares two worktrees to determine which solution better satisfies project guidelines, applying rules from contribution files to compute scores and deliver a formal verdict.

Can I use project rule files like CONTRIBUTING.md and linter configs to score code comparisons?

Yes, the analysis requires access to project rule files such as CLAUDE.md, AGENTS.md, CONTRIBUTING.md, linter configs, and CI configs to apply rules and compute scores for the worktree comparison.

What's the best way to evaluate code quality and adherence to contribution guidelines across different branches?

Providing two absolute paths to the worktrees allows the system to generate a detailed, evidence-based verdict that surfaces strengths, weaknesses, and trade-offs across correctness, design, testing, and compliance dimensions.

Do I need any dependencies or external tools to run a side-by-side worktree comparison?

No dependencies are required to run the side-by-side worktree comparison; the evaluation workflow only needs access to the two worktrees and their corresponding project rule files to apply rules and compute scores.